View Recording: Accelerating Manufacturing with AI

Nathan Lasnoski Welcome. 0:0:1.140 –> 0:0:4.750 Nathan Lasnoski We are here to talk about AI and manufacturing. 0:0:4.820 –> 0:0:9.990 Nathan Lasnoski I’m Maple Lesneski from concurrency and we’re gonna go through a lot of use cases today. 0:0:10.380 –> 0:0:11.950 Nathan Lasnoski This has been an awesome year. 0:0:12.20 –> 0:0:18.700 Nathan Lasnoski The amount of momentum that exists behind AI making a difference in manufacturing is palpable. 0:0:19.400 –> 0:0:27.910 Nathan Lasnoski Just this week, this is like the I think at least the 10th thing that I’ve done in terms of meeting with the Community on this topic from hundreds of people to this. 0:0:27.920 –> 0:0:29.10 Nathan Lasnoski It’s like it’s awesome. 0:0:29.20 –> 0:0:33.80 Nathan Lasnoski So we’re gonna be talking about AI and manufacturing and how you can make a difference in it. 0:0:33.430 –> 0:0:39.180 Nathan Lasnoski What a opportunity we have right now to like truly connect technology to the mission of our business. 0:0:39.190 –> 0:0:42.940 Nathan Lasnoski So hang on to your seats and ask questions. 0:0:42.950 –> 0:0:54.560 Nathan Lasnoski Throughout this, in the chat, Amy might be a little help me a little bit with some prompts if I have those and I’ll try my best to answer them either throughout or during the at the end of the conversation. 0:0:54.570 –> 0:0:59.180 Nathan Lasnoski So get those questions ready and happy to dig into any level of depth that you want. 0:0:59.750 –> 0:1:0.630 Nathan Lasnoski And go from there. 0:1:2.70 –> 0:1:3.320 Nathan Lasnoski So alright. 0:1:3.330 –> 0:1:3.640 Nathan Lasnoski What? 0:1:3.650 –> 0:1:5.40 Nathan Lasnoski What is happening out there? 0:1:5.170 –> 0:1:9.980 Nathan Lasnoski Well, in order to know what’s happening out there, we have to start by understanding what happened in the past. 0:1:10.210 –> 0:1:15.200 Nathan Lasnoski And if you think about the past, we sometimes measure that in human impact. 0:1:15.210 –> 0:1:20.420 Nathan Lasnoski We measure it in production, but what this shows is the impact on GDP. 0:1:20.510 –> 0:1:36.120 Nathan Lasnoski So if you think about different inventions that have happened over time that have impacted the manufacturing industry, you might be surprised to see that things like the printing press or steam engines or the Telegraph didn’t do a lot for actually increasing GDP in inflation adjusted dollars. 0:1:36.130 –> 0:1:37.670 Nathan Lasnoski It’s sort of surprising to what you thinks like. 0:1:37.680 –> 0:1:37.850 Nathan Lasnoski Whoa. 0:1:37.860 –> 0:1:46.290 Nathan Lasnoski Like kind of thought, those would have like, maybe impacted us more in the context of production, but it really didn’t impact us in the society in a dramatic way. 0:1:46.360 –> 0:1:47.820 Nathan Lasnoski But not in the context of production. 0:1:48.770 –> 0:2:3.530 Nathan Lasnoski If you look, however, at the advent of the mass production on the assembly line and transportation, you see dramatic changes in how we’re able to produce more goods and services for our customers that are aligned to their needs. 0:2:3.720 –> 0:2:10.510 Nathan Lasnoski But what you also saw during those moments was a dramatic change in the way that people engage within those production processes. 0:2:10.650 –> 0:2:18.360 Nathan Lasnoski So if you think about what happened to people before the mass production and after mass production, the jobs change dramatically. 0:2:18.710 –> 0:2:20.170 Nathan Lasnoski People were more craftspeople. 0:2:20.180 –> 0:2:26.900 Nathan Lasnoski They’re able to produce things individually, but they also work individually produced, which led to quality problems and other things. 0:2:27.110 –> 0:2:31.80 Nathan Lasnoski You went into the mass production line and you normalize production. 0:2:31.90 –> 0:2:35.810 Nathan Lasnoski You reduced costs, but you also had people that sort of function as cogs in the wheel. 0:2:36.770 –> 0:2:38.140 Nathan Lasnoski He also moved into flight. 0:2:38.270 –> 0:2:41.620 Nathan Lasnoski He moved into pattern where you could ship and produce things all over the world. 0:2:41.670 –> 0:2:55.440 Nathan Lasnoski There’s good things and bad things about that, but it changed the way that we produced within our society and the way we ship things and deliver things to customers and just the way that we were able to acquire goods then in my lifetime. 0:2:55.710 –> 0:2:57.600 Nathan Lasnoski Sorry I wasn’t around the advent of flight. 0:2:57.610 –> 0:3:7.650 Nathan Lasnoski Maybe some of you were, but in my lifetime we had the advent of the PC, the advent of Internet and the smartphone and all those changed my life. 0:3:7.820 –> 0:3:14.700 Nathan Lasnoski Thematically, I’m sure they changed yours dramatically and good ways and bad ways they they enable us to access goods and services. 0:3:14.710 –> 0:3:16.180 Nathan Lasnoski They enable us to produce more. 0:3:16.330 –> 0:3:28.320 Nathan Lasnoski They enable us to create more, but also they monopolize a lot of our life and change the way that we act in and around people change the way our interpersonal relationships work and change the way I think about myself. 0:3:28.450 –> 0:3:29.980 Nathan Lasnoski All that’s part of this picture. 0:3:29.990 –> 0:3:36.340 Nathan Lasnoski So as we go through this journey today talking about AI and manufacturing wants you to think about in two contexts. 0:3:36.710 –> 0:3:47.60 Nathan Lasnoski How do I serve the mission of my business in producing driving revenue, operational savings, but also how do I unite that to the way that I help people in my organization to make that journey? 0:3:48.260 –> 0:3:51.350 Nathan Lasnoski So we’re going to do 2 things in this conversation today. 0:3:51.360 –> 0:3:57.660 Nathan Lasnoski The first thing we’re going to do is frame up the journey and some of this is you may have seen before. 0:3:57.720 –> 0:3:59.190 Nathan Lasnoski So you may not have seen before. 0:3:59.320 –> 0:4:11.930 Nathan Lasnoski This is a really well like contextualized way in the manufacturing space to be able to show how manufacturers are making their way through the AI journey and how they’re taking advantage of these capabilities in their business. 0:4:12.120 –> 0:4:16.250 Nathan Lasnoski Then I’m going to show you companies that are taking advantage of these capabilities today. 0:4:16.260 –> 0:4:17.50 Nathan Lasnoski What are they doing? 0:4:17.60 –> 0:4:17.890 Nathan Lasnoski How are people built? 0:4:17.900 –> 0:4:25.230 Nathan Lasnoski Some of these opportunities that are available to them and brought them to market and I’m going to go through a multitude of different example. 0:4:25.240 –> 0:4:27.210 Nathan Lasnoski So keep your attention taught. 0:4:27.260 –> 0:4:30.920 Nathan Lasnoski We’re gonna go through the lots of different things, and then I’ll show you how to take action. 0:4:32.230 –> 0:4:38.320 Nathan Lasnoski So the first thing I want you to think about is that there’s really 2 domains for you to consider in the context of Manufacturing. 0:4:38.570 –> 0:4:42.280 Nathan Lasnoski The first is there’s commodity and then there’s mission driven. 0:4:42.500 –> 0:4:45.160 Nathan Lasnoski There’s commodity is everything. 0:4:45.170 –> 0:4:51.220 Nathan Lasnoski Every person within your organization being listed up to be able to create more think about it as the spell check of AI, right? 0:4:51.230 –> 0:4:53.300 Nathan Lasnoski So what would you do without spell check? Right now? 0:4:53.310 –> 0:4:56.540 Nathan Lasnoski You would have a the saurus and a dictionary sitting next to your computer. 0:4:56.550 –> 0:5:14.20 Nathan Lasnoski When you wrote emails just so you could spell things right, and you probably still wouldn’t spell things right, I know I wouldn’t, but it’s it’s a very, very sort of minimal version of what AI can look like in a sense that I’ve forgotten that I even use it. 0:5:14.110 –> 0:5:19.40 Nathan Lasnoski Commodity AI is this idea that everything I do can be done faster. 0:5:19.110 –> 0:5:37.20 Nathan Lasnoski When I start to enable the human person with ability to delegate or leverage AI tools that create value in the world, so commodity AI such as Microsoft, M365, copilot are areas where we can create opportunities for each person to be able to leverage AI as an assistant to them. 0:5:37.30 –> 0:5:40.210 Nathan Lasnoski We’ll talk about what that means and I’ll show you a couple of examples. 0:5:40.560 –> 0:5:49.70 Nathan Lasnoski The second domain is mission driven and this is truly where we are creating company specific opportunities to drive value. 0:5:49.340 –> 0:6:0.400 Nathan Lasnoski Either create a new revenue or operational savings and building opportunity within the context of accounts we’re driving these opportunities because it allows a company to execute on their strategic plan. 0:6:0.550 –> 0:6:4.800 Nathan Lasnoski You don’t create mission driven opportunities to not meet your strategic plan. 0:6:4.810 –> 0:6:9.610 Nathan Lasnoski You build a strategic plan and then you leverage those to be able to create that goods within the world. 0:6:9.690 –> 0:6:11.830 Nathan Lasnoski So we’ll show you examples of those as well. 0:6:12.960 –> 0:6:21.750 Nathan Lasnoski But before I do that, I want to hit common concerns that exist as manufacturers have been adapting and adopting AI capabilities. 0:6:22.120 –> 0:6:27.750 Nathan Lasnoski Each of these are ones that most manufacturers are running into, and ones that there’s resolutions to. 0:6:28.100 –> 0:6:31.110 Nathan Lasnoski So the first of these is this idea around data privacy. 0:6:31.180 –> 0:6:37.870 Nathan Lasnoski Most companies come to the table and say wait like once they start using AI does this all of a sudden start becoming everybody’s data? 0:6:37.920 –> 0:6:49.310 Nathan Lasnoski Like is my Manufacturing formula the schema the data that’s centric to meme all of a sudden becomes something everybody else has access to, and there’s good reason to care about that. 0:6:49.320 –> 0:6:56.140 Nathan Lasnoski Because if you’re using chat GPT publicly or you’re putting things into Google, all of a sudden it becomes something that’s marketed to you. 0:6:56.230 –> 0:7:5.820 Nathan Lasnoski Everybody’s had the experience of typing something in Google and then being like, look, I don’t need to know about this Disney vacation like, I was just checking out or where someone else was going. 0:7:5.830 –> 0:7:8.940 Nathan Lasnoski Stop trying to market me the Disney vacation, please. 0:7:9.90 –> 0:7:11.840 Nathan Lasnoski Like this data privacy concern is an issue. 0:7:12.150 –> 0:7:18.970 Nathan Lasnoski However, everything we talked about today is using a private instance, so we’re using uh. 0:7:18.980 –> 0:7:24.870 Nathan Lasnoski We’re using private instances living on Azure M365 where your data is your data. 0:7:24.880 –> 0:7:30.80 Nathan Lasnoski It is no one else’s data but yours, and that’s a really important thing to be able to convey across the organization. 0:7:31.160 –> 0:7:33.440 Nathan Lasnoski The second thing is data readiness. 0:7:33.700 –> 0:7:35.630 Nathan Lasnoski Data readiness is not a monolithic thing. 0:7:35.640 –> 0:7:49.490 Nathan Lasnoski Sometimes people use it as a cop out, like my data is not ready like your data may not be ready for certain use cases, and if it’s not, that gives us a ability to go chart the course to be able to accomplish getting the data ready for that use case. 0:7:49.640 –> 0:7:58.500 Nathan Lasnoski So we start with use case first focus on the value and then map readiness of data against that value you’re trying to drive, so you can intentionally work toward a goal. 0:8:0.420 –> 0:8:1.450 Nathan Lasnoski Human displacement. 0:8:1.460 –> 0:8:3.330 Nathan Lasnoski We started to hit on that at the beginning. 0:8:3.340 –> 0:8:6.870 Nathan Lasnoski I’m going to talk about that a little bit more of the tail end of this conversation. 0:8:7.400 –> 0:8:8.910 Nathan Lasnoski AI is a force multiplier. 0:8:9.280 –> 0:8:14.120 Nathan Lasnoski This is a rupture itself is an old topic, as we showed earlier. 0:8:14.260 –> 0:8:21.650 Nathan Lasnoski You see people go through technology transitions a lot in their lifetime and across the ages that we’ve been in existence. 0:8:21.780 –> 0:8:30.890 Nathan Lasnoski This is an opportunity for us to train and engage people in a really significant transformation, but one that is something that’s tangible to people and they could take advantage of. 0:8:30.900 –> 0:8:31.870 Nathan Lasnoski So we’ll talk more about that. 0:8:32.970 –> 0:8:38.80 Nathan Lasnoski The second biggest thing that most companies run into is things like quality and hallucination. 0:8:38.90 –> 0:8:39.660 Nathan Lasnoski Like what if the AI gets it wrong? 0:8:39.670 –> 0:8:41.300 Nathan Lasnoski What do I do about that problem? 0:8:41.650 –> 0:8:45.800 Nathan Lasnoski That is the basis of testability of applications. 0:8:45.850 –> 0:8:52.870 Nathan Lasnoski It’s what’s necessary for us to build as we build platforms for us to be able to measure it and know that it’s successful. 0:8:53.70 –> 0:9:0.120 Nathan Lasnoski Interestingly enough, what I find is that most things we build AI around when they’re only human processes they weren’t measuring in the 1st place. 0:9:0.350 –> 0:9:2.890 Nathan Lasnoski They didn’t know if the humans were answering the questions right. 0:9:3.60 –> 0:9:5.430 Nathan Lasnoski They didn’t know if their demand forecast was accurate. 0:9:5.440 –> 0:9:7.390 Nathan Lasnoski They are not measuring it against reality. 0:9:7.700 –> 0:9:22.250 Nathan Lasnoski This is an opportunity to be able to build that into everything that we do and then to validate that the answers that are being produced are accurate by building in hallucination and protection by building in measurability by testing it against and continuously improving our output. 0:9:22.840 –> 0:9:24.850 Nathan Lasnoski And then finally, the ideas around bias. 0:9:24.860 –> 0:9:44.860 Nathan Lasnoski This comes in and all sorts of different places that you wouldn’t expect, but it’s a very necessary aspect of understanding and really what this comes down to is data understanding the data that you’re using, the use cases you’re using it for, and then it’s being implemented a way that has the necessary guardrails and controls to be able to outcome in the right kinds of ways. 0:9:45.0 –> 0:9:47.990 Nathan Lasnoski So we have to manage and understand our end in mind in a sense. 0:9:49.950 –> 0:10:0.0 Nathan Lasnoski So truly going to the idea of an end in mind where we want companies to land after they go through things like envisioning and planning is a value analysis. 0:10:0.10 –> 0:10:7.960 Nathan Lasnoski We want them to know how does the mission of my business and how does AI serve it and what scenarios help me accomplish that mission. 0:10:8.130 –> 0:10:17.890 Nathan Lasnoski So these are examples of those value opportunities that exist within businesses to be able to drive that kind of opportunity within the organization. 0:10:17.900 –> 0:10:19.600 Nathan Lasnoski You’re articulating them. 0:10:19.920 –> 0:10:22.400 Nathan Lasnoski You’re selecting them, you’re putting them into categories. 0:10:22.410 –> 0:10:25.320 Nathan Lasnoski You’re determining if it’s operational savings or revenue or both. 0:10:25.330 –> 0:10:30.570 Nathan Lasnoski You’re figuring out how hard it is, and sometimes the thing that your moon shot is in the first thing you do. 0:10:30.580 –> 0:10:38.170 Nathan Lasnoski Maybe you just wanna get the rocket off the ground and validate it gets into the atmosphere like you’re doing something simple to do something hard. 0:10:38.220 –> 0:10:43.340 Nathan Lasnoski Sometimes the hard thing is what you start on readily, but you do it in incremental steps. 0:10:43.440 –> 0:10:51.470 Nathan Lasnoski This is all the things you have to plan through to be able to get to a point where you’re driving value but the value exists, you just have to open up the box and figure it out. 0:10:52.610 –> 0:11:0.400 Nathan Lasnoski So we’re gonna talk through some examples now, and I’m gonna show commodity examples 1st and then I’m gonna go into mission driven examples. 0:11:1.410 –> 0:11:4.580 Nathan Lasnoski So some examples of commodity AI for every person. 0:11:4.590 –> 0:11:5.560 Nathan Lasnoski What does this mean? 0:11:5.870 –> 0:11:10.940 Nathan Lasnoski Finding answers that you have questions to summarizing meetings, probably my favorite feature. 0:11:10.950 –> 0:11:16.820 Nathan Lasnoski I’ll show you that later doing creative work like getting me started doing something, writing that first word. 0:11:17.90 –> 0:11:28.430 Nathan Lasnoski Analytical activities like Excel copilot when you’re stacking things and grouping them and figuring out pivots planning the day, you just return from vacation and you’re saying, hey, what happened while I was gone? 0:11:28.440 –> 0:11:44.810 Nathan Lasnoski What meetings that got scheduled that I need to be aware of what’s happening in this rest of this week and I missing Amy or has this meeting been scheduled with the customer, all things that you can ask copilot to answer for you or you’ve been had been tasks that you’re performing regularly that you want to delegate to an AI agent. 0:11:46.850 –> 0:11:50.480 Nathan Lasnoski So one of my colleagues has a term never create a draft. 0:11:50.490 –> 0:11:57.30 Nathan Lasnoski Again, this is this idea of having copilot be it asset to start that first word. 0:11:57.40 –> 0:11:58.780 Nathan Lasnoski It’s always the for hardest one to write. 0:11:58.850 –> 0:12:0.810 Nathan Lasnoski Get you down the road now. 0:12:0.860 –> 0:12:9.200 Nathan Lasnoski It’s interesting this particular space, because we all know how people that use chat GPT to write their blogs for them or their LinkedIn for them. 0:12:9.210 –> 0:12:17.250 Nathan Lasnoski And I kind of loathed it like in a sense, because you can always kind of tell that they did that and there’s a sense of, like, authenticity that you want in these things. 0:12:17.590 –> 0:12:22.340 Nathan Lasnoski But there’s also a lot of things you do where you’re like, screw authenticity. 0:12:22.420 –> 0:12:24.730 Nathan Lasnoski Like I just want a statement of work written. 0:12:24.800 –> 0:12:27.590 Nathan Lasnoski I want a functional document written on our policy. 0:12:27.700 –> 0:12:32.810 Nathan Lasnoski I wanna articulate when you should leave the office or stay come in the office or how what you do when you leave the show off. 0:12:32.820 –> 0:12:41.450 Nathan Lasnoski All lights like this isn’t something I need to be particularly authentic about, like I just need a document that sets it, and using copilot to be able to do that. 0:12:41.460 –> 0:12:48.810 Nathan Lasnoski Drafting is a great time saver and a great opportunity for us to get down the road and then even things that you are just trying to do that are very authentic. 0:12:48.820 –> 0:12:51.340 Nathan Lasnoski Maybe I just need some ideas to prompt me, get me off the ground. 0:12:51.890 –> 0:12:54.170 Nathan Lasnoski This copilot is great way to kind of start down that journey. 0:12:55.260 –> 0:13:1.430 Nathan Lasnoski Same kind of idea in the context of presentations being able to gather facts that you need in those presentations. 0:13:1.440 –> 0:13:2.390 Nathan Lasnoski Get you started? 0:13:2.520 –> 0:13:6.990 Nathan Lasnoski One of my colleagues, Brian, likes to use copilot to like compare architectures. 0:13:7.0 –> 0:13:8.810 Nathan Lasnoski I’m already making a slide for it. 0:13:8.980 –> 0:13:12.790 Nathan Lasnoski Maybe I just want someone to start me on like, what’s the difference between these two architectures? 0:13:12.880 –> 0:13:14.850 Nathan Lasnoski And then I can start editing as I go. 0:13:15.40 –> 0:13:17.270 Nathan Lasnoski Is it going to be the most beautiful slide for remain? 0:13:17.280 –> 0:13:18.330 Nathan Lasnoski My experience right now? 0:13:18.340 –> 0:13:21.370 Nathan Lasnoski No, but it is a great way for you to get started. 0:13:21.430 –> 0:13:24.940 Nathan Lasnoski So you don’t have to think about every single prompt initially on your own. 0:13:27.270 –> 0:13:34.240 Nathan Lasnoski Another example that I think is just this is by far my favorite thing I’m using with copilot right now is this idea of meeting summaries. 0:13:34.330 –> 0:13:38.470 Nathan Lasnoski The idea of what happened in the meeting, what are the notes that occurred? 0:13:38.630 –> 0:13:43.600 Nathan Lasnoski What tasks happened and then being able to ask questions of the meeting? 0:13:43.610 –> 0:13:48.380 Nathan Lasnoski Like, what was the tone of the meeting like, that’s something that you normally ask a human, right? 0:13:48.390 –> 0:13:49.720 Nathan Lasnoski Like, how did it go? 0:13:49.890 –> 0:13:50.310 Nathan Lasnoski Question. 0:13:50.320 –> 0:13:52.140 Nathan Lasnoski We always ask me like, how did that meeting go? 0:13:52.290 –> 0:13:53.430 Nathan Lasnoski What was the result? 0:13:53.490 –> 0:13:57.670 Nathan Lasnoski I’m always asking my colleagues to answer questions about meetings that I wasn’t in like. 0:13:57.720 –> 0:14:6.890 Nathan Lasnoski Now I can ask copilot turn on meeting transcription, turn on recording, and I can ask copilot about the meeting that they were in to be able to give me a summary to know. 0:14:6.960 –> 0:14:8.660 Nathan Lasnoski Like what occurred, how did it go? 0:14:8.670 –> 0:14:9.820 Nathan Lasnoski What were the action items? 0:14:9.830 –> 0:14:11.280 Nathan Lasnoski Anything I need to be aware of? 0:14:11.570 –> 0:14:15.320 Nathan Lasnoski Something I use all the time now is like people just schedule me back to back. 0:14:15.390 –> 0:14:23.100 Nathan Lasnoski So I’ll join meetings and be like, tell me what happened in the last five minutes, like, and it’s funny because you always see, like, people joked about their dogs. 0:14:23.250 –> 0:14:35.620 Nathan Lasnoski And then you see the things that you actually missed in the meeting and it’s great because you get to this position where like you actually have some functional form movement and you know you can like now participate in the meeting without people having to catch you up. 0:14:36.670 –> 0:14:41.780 Nathan Lasnoski Even the ideas around tone like as you can see in that text there, like the tone is professional and collaborative. 0:14:41.890 –> 0:14:45.80 Nathan Lasnoski They did not have to resolve any conflicts like that really. 0:14:45.90 –> 0:14:47.980 Nathan Lasnoski Hit the nail on the head for me, like I didn’t have to have anything else. 0:14:48.70 –> 0:14:51.140 Nathan Lasnoski So I find this to be an imperative feature. 0:14:51.150 –> 0:14:52.140 Nathan Lasnoski I hate taking notes. 0:14:52.150 –> 0:14:55.660 Nathan Lasnoski The function of something taking notes for me is fantastic. 0:14:56.720 –> 0:15:1.170 Nathan Lasnoski Umm, when we learned from earliest copilot users, people don’t wanna give it up. 0:15:1.440 –> 0:15:3.10 Nathan Lasnoski That’s really the net net of it. 0:15:3.260 –> 0:15:7.70 Nathan Lasnoski What I’ve found is that practical copilot user is there’s certain things I love. 0:15:7.80 –> 0:15:11.310 Nathan Lasnoski There’s certain things that are just OK and there’s certain things that like just aren’t totally baked yet. 0:15:11.500 –> 0:15:13.770 Nathan Lasnoski And that’s just the reality of something like this. 0:15:13.980 –> 0:15:15.550 Nathan Lasnoski It is a true copilot. 0:15:15.560 –> 0:15:17.220 Nathan Lasnoski Is not an autopilot yet. 0:15:17.400 –> 0:15:23.730 Nathan Lasnoski It’s not a true autonomous agent that’s gonna go out and be a 10X multiplier yet, but it will be. 0:15:23.880 –> 0:15:28.860 Nathan Lasnoski And right now it’s maybe a 1.5 X multiplier, which is still a huge burst in a benefit. 0:15:29.30 –> 0:15:38.760 Nathan Lasnoski If you wanna learn more about like things I love about copilot and things I don’t love about copilot, check out my AI newsletter from 2 weeks ago where I hit that topic. 0:15:38.890 –> 0:15:42.740 Nathan Lasnoski So you can find that on LinkedIn and get some more content on that. 0:15:44.300 –> 0:15:44.750 Nathan Lasnoski OK. 0:15:45.40 –> 0:15:49.210 Nathan Lasnoski Umm, so we could go into like copilot use cases and commodity use cases all day. 0:15:49.220 –> 0:16:0.590 Nathan Lasnoski Uh, I guess one more thing on commodity is that copilot is not the only example of commodity that you’re gonna find like realize that like AI capabilities are gonna light up in most of the platforms that you use. 0:16:0.700 –> 0:16:5.490 Nathan Lasnoski And you’re gonna dock them as an AI practitioner in a lot of things that you’re taking advantage of. 0:16:5.810 –> 0:16:9.820 Nathan Lasnoski OK, so now we’re going to go through a multitude of. 0:16:11.700 –> 0:16:20.50 Nathan Lasnoski Scenarios being used in customers today in ways that they’ve taken advantage of us to, to build value and their direct organizations. 0:16:21.330 –> 0:16:27.800 Nathan Lasnoski So things you need to think about in this space if you are thinking about GPT driven models, there’s a lot of value here. 0:16:28.10 –> 0:16:37.600 Nathan Lasnoski Things like service request automation, any kind of customer service scenario, huge opportunity for GPT driven conversations to drive value service status. 0:16:37.610 –> 0:16:38.640 Nathan Lasnoski Where’s my truck? 0:16:38.650 –> 0:16:41.380 Nathan Lasnoski Where’s my thing? Interaction acceleration? 0:16:41.390 –> 0:16:53.790 Nathan Lasnoski Like quoting one of the our bigger case studies is driving quoting down driving, quoting into a level where like any kind of course you produce can be delivered to a customer faster. 0:16:54.160 –> 0:16:57.740 Nathan Lasnoski Data mining and insights asking questions of your business system. 0:16:57.750 –> 0:16:58.750 Nathan Lasnoski What’s my lead time? 0:16:58.760 –> 0:17:0.250 Nathan Lasnoski What’s my inventory backlog? 0:17:0.260 –> 0:17:1.900 Nathan Lasnoski When is this product being delivered? 0:17:2.760 –> 0:17:13.670 Nathan Lasnoski Especially, will it customers on a call or do having to do so in an analytic way and then anything you do a lot like any kind of processing that you have a person do a hand processing thing. 0:17:13.860 –> 0:17:16.600 Nathan Lasnoski These are a great opportunities for us to look at for AI. 0:17:16.840 –> 0:17:19.970 Nathan Lasnoski And then note that this is just a piece of it. 0:17:20.460 –> 0:17:29.110 Nathan Lasnoski Lot of the bigger that’s use cases are supply chain automation and predictive maintenance, smart warehouse vision oriented scenarios. 0:17:29.340 –> 0:17:37.390 Nathan Lasnoski There’s a ton that’s not truly generative AI, or maybe combines with it to be able to create some really amazing outcomes. 0:17:37.680 –> 0:17:39.210 Nathan Lasnoski So let’s go through some examples. 0:17:39.220 –> 0:17:47.380 Nathan Lasnoski So this is one starts to give you a picture of how you connect some of the pieces to do things like customer service, Oregon sales enablement. 0:17:47.710 –> 0:17:51.680 Nathan Lasnoski So you can see for example, I’m asking questions of this boat. 0:17:52.530 –> 0:17:53.570 Nathan Lasnoski What does the draft? 0:17:53.580 –> 0:18:10.750 Nathan Lasnoski What’s the maximum speed which you can see in this is I have a speed, it’s found in a reference that is this document right here you can see it’s even understanding the draft equals the measurement, because even in the document doesn’t say it necessarily what the draft is. 0:18:11.280 –> 0:18:21.850 Nathan Lasnoski There’s a measurement which is essentially how low can go in the how low in the water it goes in conjunction with the depth of the water, and you can see they’re citation that it’s pulling me from. 0:18:21.960 –> 0:18:30.70 Nathan Lasnoski So this really kind of gets across some things that we do to a mitigate hallucination, this idea that, like it’s answering a question, but is it trustworthy? 0:18:30.500 –> 0:18:35.30 Nathan Lasnoski Yes, because we’re showing a citation, we’re referencing that it got it from that citation. 0:18:35.40 –> 0:18:42.980 Nathan Lasnoski We’re giving you the ability to go there and look for yourself, so if the person has a question about it, they can actually dig in more or ask more questions of the bot. 0:18:43.670 –> 0:18:54.640 Nathan Lasnoski I find this to be a fantastic way for for companies to kind of get their feet wet in the customer service base by arming their customer service reps or their sales people with information so they can help their customers faster. 0:18:54.870 –> 0:19:9.960 Nathan Lasnoski Think about this also in the context of like even internal use cases, HR bots answering questions about your company policy or your sick leave things that like you would answer internally for your own employees can be done as a great initial use case as well. 0:19:11.320 –> 0:19:29.620 Nathan Lasnoski This example for a company Generac they’re one of the larger power standby power producers in the nation and one of the things that they looked at using is and a customer service bot to help their customer service reps answer questions about about questions on the generator. 0:19:29.630 –> 0:19:30.290 Nathan Lasnoski So they call in. 0:19:30.300 –> 0:19:33.0 Nathan Lasnoski They say, hey, how do I do this thing like standby power down? 0:19:33.10 –> 0:19:44.240 Nathan Lasnoski My sent hummus home standby generator and I give them a process that they use to be able to take that action as well as the source, and did it help you or did it help you? 0:19:44.310 –> 0:20:1.830 Nathan Lasnoski So this opportunity to be able to create that value while that customers on the call reducing the time to serve that question and delivering them an answer faster and then moving that to even a direct customer conversation, I don’t even need the customer service Rep necessarily answering every question. 0:20:1.960 –> 0:20:6.400 Nathan Lasnoski Maybe there’s certain questions that AI bought could do on its own, and you’re even seeing. 0:20:6.410 –> 0:20:19.10 Nathan Lasnoski I don’t have this example in here, but if you’re doing your taxes this year and this is in a Manufacturing example, but I think it’s a good one if you’re doing your taxes here and you’re using TurboTax, check out their bot that’s on the side of your tax process. 0:20:19.260 –> 0:20:33.60 Nathan Lasnoski And it’s very it’s like sudo Interactive definitely injects into play areas that you missed filling out your taxes and you can see how even like financial services providers that have pretty high bars are building some of these chat experiences into their products. 0:20:34.390 –> 0:20:40.180 Nathan Lasnoski Umm, so moving down the road, one of the things that you might also find interesting is things like defect detection. 0:20:40.370 –> 0:20:48.880 Nathan Lasnoski So this is an example of very simple example and then again do a more complicated one later about comparing a good against a bad. 0:20:49.30 –> 0:20:53.580 Nathan Lasnoski So this is a good example of a screw. 0:20:53.650 –> 0:20:57.680 Nathan Lasnoski As we can see, and then as all of us have experienced, we’ve stripped screws in our lifetime. 0:20:57.890 –> 0:21:0.880 Nathan Lasnoski These are some not good screws that we are comparing against. 0:21:1.70 –> 0:21:4.770 Nathan Lasnoski Doesn’t know how to correctly defer determine if that’s a good screw or not. 0:21:5.140 –> 0:21:5.790 Nathan Lasnoski It does. 0:21:5.960 –> 0:21:7.680 Nathan Lasnoski So this is screws effective. 0:21:7.690 –> 0:21:10.150 Nathan Lasnoski This is not defective, defective, defective, right? 0:21:10.160 –> 0:21:16.110 Nathan Lasnoski So this is using a GPT model that had very little need for actual training. 0:21:16.120 –> 0:21:16.290 Nathan Lasnoski What? 0:21:16.300 –> 0:21:21.990 Nathan Lasnoski One of the things I think is really interesting about this is typically to do that I have to give it many examples. 0:21:22.120 –> 0:21:29.890 Nathan Lasnoski Like many, many, many examples of what a bad in like, a bad good, I have to give it like 100 and 100, right? 0:21:30.140 –> 0:21:44.820 Nathan Lasnoski Or even more than that, with things like GPT 4V we’re seeing that I can do a lot less work to be able to get to that same outcome where I could feed it a manual that shows the inside of a cockpit of a a boat or a car or something. 0:21:44.930 –> 0:21:56.500 Nathan Lasnoski And then it can correctly discern what’s happening within the context of that to even give feedback back in that manual use case where in the past, even early last year, we couldn’t do that because it was only text based. 0:21:56.570 –> 0:22:4.30 Nathan Lasnoski Now we’re getting to a point where we can combine text with the image and ask really intelligent questions that haven’t answer it in an intelligent way. 0:22:6.950 –> 0:22:23.100 Nathan Lasnoski So continuing down the road of valuable use cases in the manufacturing space, I want you to think about this idea around demand forecasting, demand, inventory, supply chain relationships with your customers, all that’s interconnected sort of spaghetti weave, right? 0:22:23.110 –> 0:22:24.820 Nathan Lasnoski We have all that stuff together. 0:22:25.190 –> 0:22:37.810 Nathan Lasnoski This is an example of a company that used AI to be able to optimize their demand and inventory forecasting process across multiple locations and multiple countries to be able to optimize their inventory on hand. 0:22:37.880 –> 0:22:44.90 Nathan Lasnoski So they had to spend less money to keep inventory that they weren’t using and applied the right money to the right places. 0:22:44.200 –> 0:22:50.830 Nathan Lasnoski So they saved between 5 and 20% between location, freeing up capital of 50 to 80 million year over year. 0:22:50.940 –> 0:23:0.10 Nathan Lasnoski In this case, study goes back to 2018, so this is something companies have been doing for a long, long time. 0:23:0.160 –> 0:23:5.230 Nathan Lasnoski What’s changed is the accessibility of these projects like this was $1,000,000 project. 0:23:5.920 –> 0:23:18.800 Nathan Lasnoski Dropping that down now into the like the 10s or hundreds of thousands of dollars gets these things to a position where companies can realistically take action on them and drive value in their organizations. 0:23:18.810 –> 0:23:24.680 Nathan Lasnoski That takes effect faster and a lot of this is already packaged like this is something we have an accelerator for. 0:23:24.690 –> 0:23:26.170 Nathan Lasnoski We can go do it tomorrow. 0:23:26.330 –> 0:23:31.610 Nathan Lasnoski These are things that we can take in, accelerate and bring value on faster than we could say in 2018. 0:23:32.910 –> 0:23:45.80 Nathan Lasnoski Same kind of idea with quoting, taking a quote from a customer or a quest that comes through email, pulling out relevant information, accepting or modifying it, and sending it back to customers. 0:23:45.90 –> 0:23:49.40 Nathan Lasnoski We worked with in this space, have dropped their time to quote from hours down to minutes. 0:23:49.130 –> 0:23:57.600 Nathan Lasnoski And the reason why is because taking writing up a quote takes time, taking information, interpreting it, comparing against your available products takes time. 0:23:57.830 –> 0:24:2.100 Nathan Lasnoski Developing that into something that a customer’s gonna accept and is the right price. 0:24:2.270 –> 0:24:7.870 Nathan Lasnoski Sometimes you don’t even know the person is even know what to price it, and that’s even an opportunity to be able to do pricing strategy. 0:24:8.170 –> 0:24:16.720 Nathan Lasnoski So on the sales side, there’s traumatic opportunity to be able to increase revenue by engaging customers with new opportunities in this space. 0:24:16.730 –> 0:24:17.860 Nathan Lasnoski So, awesome. 0:24:17.870 –> 0:24:29.180 Nathan Lasnoski Awesome opportunities, especially in the quoting space, to be able to create more revenue by being faster or more targeted or aligned with your customers or even increasing margin by using the right pricing for the type of scenario. 0:24:30.850 –> 0:24:36.210 Nathan Lasnoski So that you that’s that example I showed you was relatively simple example, right? 0:24:36.220 –> 0:24:39.150 Nathan Lasnoski Like it was like, hey, I here’s some stuff in the email. 0:24:39.640 –> 0:24:41.10 Nathan Lasnoski Now I’m putting it in. 0:24:41.20 –> 0:24:42.140 Nathan Lasnoski Now I’m creating a quote right? 0:24:42.150 –> 0:24:44.160 Nathan Lasnoski Like it’s more commoditized. 0:24:45.290 –> 0:24:47.140 Nathan Lasnoski Let’s think about a more custom scenario. 0:24:47.530 –> 0:24:55.400 Nathan Lasnoski Window manufacturer needs to think about how do I do custom windows that are a huge part of My Portfolio. 0:24:55.690 –> 0:24:57.840 Nathan Lasnoski That’s where that data readiness comes into the picture. 0:24:57.850 –> 0:25:2.340 Nathan Lasnoski I need a schema that sits underneath that that helps me to build my end in mind. 0:25:2.530 –> 0:25:11.50 Nathan Lasnoski So if you’re end in mind, is this then anybody bought a window lately and you’ve gone to Home Depot and you’ve been sitting there with the person they bring this book out and you’re like, paging through. 0:25:11.60 –> 0:25:12.530 Nathan Lasnoski And it’s like, it’s like a terrible experience. 0:25:12.540 –> 0:25:14.360 Nathan Lasnoski It’s not very not very digital. 0:25:14.500 –> 0:25:16.960 Nathan Lasnoski It’s there’s not very compared, they barely know the windows. 0:25:17.460 –> 0:25:18.900 Nathan Lasnoski What if I could take my smartphone? 0:25:18.910 –> 0:25:20.90 Nathan Lasnoski Hold it up to the window. 0:25:20.400 –> 0:25:29.730 Nathan Lasnoski It can transpose the AR over it and give me a accurate ability to be able to kind of figure out what I want for myself and then self quote. 0:25:29.880 –> 0:25:32.170 Nathan Lasnoski Well, that requires a schema underneath it. 0:25:32.230 –> 0:25:43.780 Nathan Lasnoski That’s where your data readiness is really important, but it allows you to be able to change the game in your market, and this is where you start to disrupt things that other companies are doing because you’re able to engage with customers directly. 0:25:45.0 –> 0:25:52.740 Nathan Lasnoski So, uh, fascinating opportunities for you to think about how your business goes to market with your customers and whether there’s opportunity to change that conversation. 0:25:54.620 –> 0:25:55.10 Nathan Lasnoski OK. 0:25:55.20 –> 0:25:58.230 Nathan Lasnoski So another thing, that’s man, you’re stick with me here. 0:25:58.240 –> 0:26:3.400 Nathan Lasnoski There is a lot of examples and I want you to make sure you’re seeing all these. 0:26:3.410 –> 0:26:7.750 Nathan Lasnoski So this next example is Horus Electronic. 0:26:7.760 –> 0:26:22.240 Nathan Lasnoski They’re out Washa, Wisconsin, and if anyone has seen the Terminator movies, Terminator two, you know, like when Arnold gets lowered into the VAT of molten metal and he’s got the thumbs up and stuff, it’s like when any that time. 0:26:22.250 –> 0:26:25.230 Nathan Lasnoski We’re like every man in your life is allowed to cry like. 0:26:25.240 –> 0:26:26.170 Nathan Lasnoski Yeah, that’s that’s this. 0:26:26.180 –> 0:26:27.410 Nathan Lasnoski OK, so they do this. 0:26:27.420 –> 0:26:31.740 Nathan Lasnoski They have, like, these molten metal vases, and they create these custom metal products, OK. 0:26:33.590 –> 0:26:40.100 Nathan Lasnoski That is not a safe environment for people to be like walking around and like dipping things into to test the production process. 0:26:40.470 –> 0:26:48.500 Nathan Lasnoski This is a really interesting way that they’re able to connect their OT environment to provide intelligent guidance of how they did their production. 0:26:48.610 –> 0:26:55.820 Nathan Lasnoski So they can reduce the cost of waste and optimize the production process to deliver the best outcome to their customers. 0:26:56.250 –> 0:27:18.270 Nathan Lasnoski So this is a space where I was at as CDIO event in one of the companies that was speaking at that event was talking about their AI use case very similar to this where like I think you refer to OT as like this like this wealth of data that we barely have even touched like it’s like don’t cross the streams and we because we all know that environments are mess. 0:27:18.280 –> 0:27:19.510 Nathan Lasnoski It’s got Windows XP in it. 0:27:19.700 –> 0:27:20.710 Nathan Lasnoski Got just, it’s horrific. 0:27:20.720 –> 0:27:21.310 Nathan Lasnoski Like, don’t. 0:27:21.600 –> 0:27:29.210 Nathan Lasnoski When you walk around your factory, just like, Oh my gosh, I can’t believe this is actually running, but realize that that data coming off the machine is super valuable. 0:27:29.540 –> 0:27:36.700 Nathan Lasnoski This this uh CDI O is saying there was part of their production process that they didn’t understand and had never understood. 0:27:36.870 –> 0:27:43.740 Nathan Lasnoski It was this diversity of outcome that happened as they did their production, or sometimes the process was like 30% efficient. 0:27:43.750 –> 0:27:47.0 Nathan Lasnoski Other times was like 60% efficient and they weren’t sure why. 0:27:47.10 –> 0:27:48.670 Nathan Lasnoski Like it just was like always diverse. 0:27:48.680 –> 0:27:50.380 Nathan Lasnoski And they had to just manage it real time. 0:27:50.870 –> 0:27:59.120 Nathan Lasnoski And what they were able to do was they took some data from that and they were able to engage data scientists to be able to figure out what that gap was. 0:28:0.150 –> 0:28:4.800 Nathan Lasnoski So in that space, data and data science and AI was a function of. 0:28:4.810 –> 0:28:10.40 Nathan Lasnoski It was like a tool to be able to discover something that they didn’t know that was not intuitive about their environment. 0:28:10.110 –> 0:28:15.700 Nathan Lasnoski That led to an optimization of how they’re doing their production, which is like holy cow, it’s awesome. 0:28:15.710 –> 0:28:25.940 Nathan Lasnoski Example because it allows you to be able to actually produce more out of the same factory because you know more about your production process that these people have been working there for 40 years could never figure out. 0:28:25.950 –> 0:28:33.310 Nathan Lasnoski So there’s some really interesting things that can be done just in the experiment in place that in the manufacturing environment that can be of value. 0:28:35.130 –> 0:28:38.270 Nathan Lasnoski OK, so uh, this is your quiz for today. 0:28:38.330 –> 0:28:38.600 Nathan Lasnoski Umm. 0:28:38.610 –> 0:28:42.580 Nathan Lasnoski I’m expecting a whole bunch of things in the chat. 0:28:43.270 –> 0:28:46.20 Nathan Lasnoski What do you think this is? 0:28:46.30 –> 0:28:51.420 Nathan Lasnoski If you had to guess what this is on the screen, what do you think that it is? 0:28:52.990 –> 0:28:53.540 Nathan Lasnoski Give you. 0:28:53.550 –> 0:28:56.370 Nathan Lasnoski I’ll give you less than, uh, less than 5 seconds here. 0:28:57.680 –> 0:28:58.930 Nathan Lasnoski What do you think it is? 0:29:1.260 –> 0:29:1.890 Nathan Lasnoski Mars. 0:29:1.900 –> 0:29:2.810 Nathan Lasnoski That’s a great guess. 0:29:4.140 –> 0:29:5.770 Nathan Lasnoski Tunnel defect. 0:29:5.880 –> 0:29:8.570 Nathan Lasnoski Chimney wound care. 0:29:8.640 –> 0:29:11.380 Nathan Lasnoski Yes, yes, sewer pipe. 0:29:11.390 –> 0:29:14.130 Nathan Lasnoski OK, defect in a bearing. 0:29:16.260 –> 0:29:19.330 Nathan Lasnoski Foundation wall looks like the lid for a pipe. 0:29:19.400 –> 0:29:20.870 Nathan Lasnoski Eyeball. Eyeball. 0:29:20.880 –> 0:29:21.200 Nathan Lasnoski Come on. 0:29:23.20 –> 0:29:23.590 Nathan Lasnoski Good guess. 0:29:23.600 –> 0:29:25.790 Nathan Lasnoski OK, I nobody said Colon. 0:29:25.840 –> 0:29:27.820 Nathan Lasnoski I was really expecting you someone to say colon. 0:29:27.830 –> 0:29:28.670 Nathan Lasnoski It kind of looks like one. 0:29:28.680 –> 0:29:31.490 Nathan Lasnoski It’s actually a sewer pipe sewer, said sewer pipe guest correctly. 0:29:31.960 –> 0:29:34.10 Nathan Lasnoski This is a company called Burgers and Nipel. 0:29:34.20 –> 0:29:37.810 Nathan Lasnoski They lower like little robots into sewer pipes to look for damage. 0:29:37.820 –> 0:29:51.310 Nathan Lasnoski Actually, before engaging these guys, I didn’t realize how much water gets lost from cracks and sewer pipes like you think runoffs, a problem like the percentage of water they actually makes it from one end of the sewer pipe to that other is dramatically low. 0:29:51.900 –> 0:29:55.660 Nathan Lasnoski So this robot goes under there and look for cracks and sewer pipes, right? 0:29:55.670 –> 0:30:1.570 Nathan Lasnoski So what AI is doing in this use case is it’s instead of the person watching the entire process. 0:30:2.0 –> 0:30:12.640 Nathan Lasnoski They’re it’s identifying issues and then pointing the person to that, that gap in the pipe and then saying here’s where it needs to be remediated. 0:30:12.650 –> 0:30:15.890 Nathan Lasnoski And here’s where it doesn’t this can be used for all sorts of types of use cases. 0:30:16.0 –> 0:30:31.160 Nathan Lasnoski Think about flying drones that look for damage, are on high tension power lines, for example, or go around your factory and do cycle counting like there’s opportunities all over the place with visioned dramatically impact your business. 0:30:31.350 –> 0:30:34.240 Nathan Lasnoski So I thought this was really cool because they’re looking for. 0:30:34.250 –> 0:30:40.130 Nathan Lasnoski They’re taking the very nature of their organization and force multiplying a person in the task that we’re doing. 0:30:41.760 –> 0:30:44.790 Nathan Lasnoski OK, umm this this example is the like. 0:30:44.800 –> 0:30:45.990 Nathan Lasnoski Where’s my stuff? 0:30:46.0 –> 0:30:52.790 Nathan Lasnoski Domino’s Pizza tracker I don’t know if people are aware of this, but like initially when Domino’s came out with the Domino’s Pizza tracker, it wasn’t real. 0:30:52.800 –> 0:30:57.610 Nathan Lasnoski Like they just kind of took the average time and put it in front of everybody, like randomizes a little bit. 0:30:57.620 –> 0:30:59.190 Nathan Lasnoski Like, here’s your Domino’s Pizza tracker. 0:30:59.760 –> 0:31:4.610 Nathan Lasnoski It’s real now, but in this context, what we’re trying to do is set expectations. 0:31:4.620 –> 0:31:12.670 Nathan Lasnoski One of the bigger issues I see with manufacturers, it’s hard for them to set expectations for their customers, especially when they have a diverse set of incoming sales and pipeline. 0:31:12.680 –> 0:31:22.40 Nathan Lasnoski And no, what they’re producing on what line and what day in which you can see here is we have a uh green light green green. 0:31:22.50 –> 0:31:25.840 Nathan Lasnoski Then we have this yellow section right, which is kind of institutes like we’re missing something. 0:31:25.850 –> 0:31:26.770 Nathan Lasnoski We’re missing apart. 0:31:26.780 –> 0:31:27.540 Nathan Lasnoski We’re missing something. 0:31:27.550 –> 0:31:30.800 Nathan Lasnoski Lets us produce and what AI does in this context? 0:31:30.810 –> 0:31:38.980 Nathan Lasnoski Is it takes the human out of the loop and inserts a lot more objectivity to when this is going to be delivered to my customer. 0:31:39.230 –> 0:31:43.380 Nathan Lasnoski A lot of times when people have like interruption in the supply chain, they have a person running around. 0:31:43.390 –> 0:31:44.560 Nathan Lasnoski People like when you’re going to get this. 0:31:44.570 –> 0:31:45.380 Nathan Lasnoski Is it coming tomorrow? 0:31:45.390 –> 0:31:45.940 Nathan Lasnoski I don’t know. 0:31:45.950 –> 0:31:46.880 Nathan Lasnoski Just tell them 10 days. 0:31:46.890 –> 0:31:49.530 Nathan Lasnoski You know, you get this, like, kind of diversity of vinyl. 0:31:49.540 –> 0:31:50.290 Nathan Lasnoski Answer them. 0:31:50.520 –> 0:31:58.470 Nathan Lasnoski This is what allows you to set some objectivity behind that to be able to be more consistent and also to measure whether you’re objectivity was objective. 0:31:58.640 –> 0:32:1.10 Nathan Lasnoski Did I correctly estimate? 0:32:1.20 –> 0:32:2.590 Nathan Lasnoski Did I have better customer outcomes? 0:32:2.600 –> 0:32:9.460 Nathan Lasnoski And ultimately, what’s trying to do customers is sell them more things, keep them happy, make sure that what you’re delivering to them is delivered on time. 0:32:9.470 –> 0:32:18.30 Nathan Lasnoski You’re matching the expectations and you might tell them 10 days and they’re OK with it as long as you didn’t tell them five days and you’re 5 days late, right? 0:32:18.40 –> 0:32:21.380 Nathan Lasnoski Like tell them 10 days from the get go and you’ll have a happy customer. 0:32:21.390 –> 0:32:29.660 Nathan Lasnoski At least you’ll have a customer that’s has expectations managed versus like you said, five days now it’s 10 days. 0:32:29.750 –> 0:32:34.180 Nathan Lasnoski I’ve just got it now that had me, I didn’t have my people ready for it to be delivered. 0:32:34.290 –> 0:32:40.380 Nathan Lasnoski You set this whole problem up, so this really comes into play and it’s great opportunity to engage AI. 0:32:40.390 –> 0:32:42.560 Nathan Lasnoski And then this can also be a query driven experience. 0:32:42.570 –> 0:32:50.590 Nathan Lasnoski It could be like when is the order being delivered for person X and then have it answer that question for you versus have to go into your business system and go find it. 0:32:52.640 –> 0:33:15.70 Nathan Lasnoski OK, so some things similar to that is this sort of process automation space and on the lot of people went down the road of RPA found that like it was helpful in some areas, but like struggled because it couldn’t do generalizable things and that generalizable space is where we’re seeing opportunities to gain in things like contract automation for example. 0:33:15.280 –> 0:33:23.710 Nathan Lasnoski So in this use case you have hundreds of thousands of contracts that come through every like purchase orders matching to sales orders, matching to production orders. 0:33:23.720 –> 0:33:28.980 Nathan Lasnoski That have to happen over a year in a person is looking at those and make sure am I making the right thing like did. 0:33:28.990 –> 0:33:32.340 Nathan Lasnoski I did what I entered match what they put in their PO. 0:33:32.350 –> 0:33:38.740 Nathan Lasnoski That what I’m charging them, that I’m ultimately gonna ship because there’s by build the wrong thing and ship it and they’re unhappy. 0:33:38.750 –> 0:33:41.800 Nathan Lasnoski Like, that’s a lot of problems like lot of problems. 0:33:41.810 –> 0:33:48.40 Nathan Lasnoski I have to now untangle so companies have people that do this, analyze the PO, match it, validate it. 0:33:48.450 –> 0:34:11.830 Nathan Lasnoski This is a good example to use as you think about our automation opportunities with AI, because 100% is not necessary for success, I could do 80% of these automated and have 20% go to a human for intervention and then incrementally go to 81 percent, 82 percent, 83% and I’m winning every single time because I’m automating that process. 0:34:12.140 –> 0:34:24.770 Nathan Lasnoski This is the opportunity to be able to create real value within your organization without having to be like if I don’t get 100% puppies die, you know this is this is an opportunity to create real value. 0:34:24.880 –> 0:34:28.530 Nathan Lasnoski So customer contract automation is is an example of that. 0:34:28.540 –> 0:34:30.470 Nathan Lasnoski It could be almost anything in your business. 0:34:31.590 –> 0:34:36.980 Nathan Lasnoski Umm, so now as you’ve gone through this I want I want you to think about the examples we’ve used. 0:34:37.90 –> 0:34:45.100 Nathan Lasnoski Most of them probably with the example of like the AR window use case where incremental examples they’re like I already do this. 0:34:45.150 –> 0:34:47.460 Nathan Lasnoski I’m just doing it faster now. 0:34:47.470 –> 0:34:53.160 Nathan Lasnoski We’re gonna talk about some examples that are in the disruptive space, like I don’t do this today. 0:34:53.170 –> 0:35:1.770 Nathan Lasnoski I don’t do it this way today and I want to transform that into a better way of doing it than engages the market differently, and that’s called disruptive innovation. 0:35:2.350 –> 0:35:11.260 Nathan Lasnoski So this first example I this is actually a video I for whatever reason it wasn’t playing so I didn’t move it to a a still image, but a guy walks around their car. 0:35:11.270 –> 0:35:26.820 Nathan Lasnoski It’s damaged with the video the upload that video into this insurance provider and then that video is then summarized through GPT 4V and correctly describes the damage that’s on the car. 0:35:27.210 –> 0:35:33.660 Nathan Lasnoski So the rear side of the blue Toyota Camry has sustained significant damage, characterized by deep scratches and scuff marks. 0:35:33.670 –> 0:35:36.370 Nathan Lasnoski These marks are the most prominent on the wheel, blah blah blah blah blah, right? 0:35:36.380 –> 0:35:40.980 Nathan Lasnoski It’s like describing the whole thing and the amount of the car, and we’re it came from. 0:35:41.520 –> 0:35:48.210 Nathan Lasnoski This is taking think about as taking a task that your team doesn’t like doing and describing it and doing it. 0:35:48.550 –> 0:35:51.70 Nathan Lasnoski Another example, this would be like your customer service calls. 0:35:51.320 –> 0:35:52.170 Nathan Lasnoski Record those guys. 0:35:52.280 –> 0:35:53.30 Nathan Lasnoski What happens? 0:35:53.290 –> 0:35:53.990 Nathan Lasnoski Record it. 0:35:54.60 –> 0:36:6.850 Nathan Lasnoski It can actually do the documentation of the customer service call in your customer service system for you real time like you saw a documented the meeting have it do everything for you in the customer service call like it can capture all that and do it. 0:36:6.940 –> 0:36:15.730 Nathan Lasnoski What about taking a video of the truck before it leaves and logistics, or taking a video of the products as they come off the assembly line? 0:36:15.740 –> 0:36:18.50 Nathan Lasnoski Is the sticker on properly as it been put together? 0:36:18.60 –> 0:36:19.590 Nathan Lasnoski Well, we had a customer yesterday. 0:36:19.600 –> 0:36:22.360 Nathan Lasnoski I was talking about installation of their product and bars. 0:36:22.830 –> 0:36:26.460 Nathan Lasnoski It was like is the Barr configuration set up right? 0:36:26.790 –> 0:36:28.180 Nathan Lasnoski Like was this done? 0:36:28.190 –> 0:36:31.820 Nathan Lasnoski Well, was the repair done well, what’s wrong with it? 0:36:31.910 –> 0:36:38.340 Nathan Lasnoski All of this is things that like just to do an AR model on a video model, and this was like $500,000 before it like. 0:36:38.350 –> 0:36:53.390 Nathan Lasnoski Now we’re getting to a point where this is very attractive because all that prep work has been done for it to understand in a sense like not only human understands, but like in a more closer to the way human understands video and image content to be able to do some really amazing things for us. 0:36:54.630 –> 0:36:56.990 Nathan Lasnoski Another space is coming to play and I use. 0:36:57.0 –> 0:37:1.410 Nathan Lasnoski This is totally on a Manufacturing use case, but I think it helps you to understand the power. 0:37:1.420 –> 0:37:2.500 Nathan Lasnoski That’s possible. 0:37:2.550 –> 0:37:9.20 Nathan Lasnoski Power the donor that the power that’s possible as a result of building these sort of enterprise class chat bots. 0:37:9.30 –> 0:37:13.460 Nathan Lasnoski OK, so self service loan application. 0:37:13.470 –> 0:37:14.600 Nathan Lasnoski Why is this powerful? 0:37:14.810 –> 0:37:24.140 Nathan Lasnoski Well, every time you do a loan, you pay about $2000 to a loan officer to do that initial credit poll associated with asking you a whole bunch of questions. 0:37:24.310 –> 0:37:28.720 Nathan Lasnoski This company is automating that process with a chat bot. 0:37:28.830 –> 0:37:36.700 Nathan Lasnoski They’re already live in California, enabling a person to do that request without the need to use any actual person. 0:37:36.770 –> 0:37:38.120 Nathan Lasnoski They’re gonna do it for free. 0:37:38.430 –> 0:37:39.240 Nathan Lasnoski Holy cow. 0:37:39.250 –> 0:37:41.700 Nathan Lasnoski Like you got tons of people that do that job. 0:37:41.750 –> 0:37:43.50 Nathan Lasnoski What’s gonna happen to their job? 0:37:43.60 –> 0:37:44.340 Nathan Lasnoski Well, that’s gonna have to change, right? 0:37:44.350 –> 0:37:49.760 Nathan Lasnoski They’re gonna have to adapt and move and do different things and add more value on top of it if they want you to pay them. 0:37:50.590 –> 0:37:51.840 Nathan Lasnoski Same thing with the realtor space. 0:37:51.850 –> 0:38:6.830 Nathan Lasnoski Honestly, like the whole space is getting interrupted, but this is really interesting because you can ask all the same questions in a really rigorous chat bot and they’re as you do your taxes, you’re going to find that like the TurboTax experience is doing the exact same thing like it’s very rigorous. 0:38:6.840 –> 0:38:8.370 Nathan Lasnoski It’s making sure you fill in all the boxes. 0:38:8.380 –> 0:38:15.970 Nathan Lasnoski It’s asking you really interesting questions and actually it’s preferable like I’d rather do that than talk to a human being in that process. 0:38:16.90 –> 0:38:20.830 Nathan Lasnoski Like, can you please just give me a robot I can interact with for certain things? 0:38:20.840 –> 0:38:22.110 Nathan Lasnoski Other things you’re like. 0:38:22.160 –> 0:38:23.230 Nathan Lasnoski I don’t want a robot. 0:38:23.280 –> 0:38:27.380 Nathan Lasnoski I wanna per human and moving that in the right use cases to a person. 0:38:27.730 –> 0:38:31.500 Nathan Lasnoski We’re partnering the two together can be a really effective pattern. 0:38:31.670 –> 0:38:40.160 Nathan Lasnoski So self service of customer service, a lot of quoting use cases is really powerful in the manufacturing and production space. 0:38:41.190 –> 0:38:44.880 Nathan Lasnoski I showed an example earlier about chat bot. 0:38:44.890 –> 0:38:46.580 Nathan Lasnoski This is sort of going to that healthcare chat. 0:38:46.590 –> 0:38:53.700 Nathan Lasnoski This is a company that is bringing a chat bot to the table associated with insurance coverage. 0:38:53.770 –> 0:38:56.430 Nathan Lasnoski So if you think about, I’ve got my insurance coverage. 0:38:56.440 –> 0:38:57.920 Nathan Lasnoski What’s the worst thing about insurance? 0:38:57.930 –> 0:39:5.980 Nathan Lasnoski Pretty much everything, but like, what’s the second most worst thing calling up customer support and like having to ask them like, is this covered in my deductible? 0:39:5.990 –> 0:39:8.380 Nathan Lasnoski And why did you deny my claim and like, please fix it? 0:39:8.390 –> 0:39:31.80 Nathan Lasnoski Because it actually is covered like all that stuff that we deal with, being able to have a chat experience, that’s asynchronous where I don’t have to go take time away from other things to be able to answer simple questions at least like what’s my deductible and is it used up already and is this thing called would be great things, I could use the chat bot for and yeah, you immediately go to like, wait, what if it gets it wrong? 0:39:31.130 –> 0:39:33.780 Nathan Lasnoski That’s where hallucination protection has to be an asset. 0:39:33.870 –> 0:39:36.480 Nathan Lasnoski But then remember, your humans are getting it wrong too. 0:39:36.750 –> 0:39:42.100 Nathan Lasnoski So this is a great opportunity for us to do that, do that well, OK. 0:39:42.250 –> 0:39:50.900 Nathan Lasnoski So the last no second to last disruptive example I want you to be aware of is what what the company Brunswick is doing in smart boating. 0:39:51.270 –> 0:39:54.100 Nathan Lasnoski They have a mission of being the leader in digital voting. 0:39:54.370 –> 0:39:55.240 Nathan Lasnoski What does that mean? 0:39:55.350 –> 0:40:0.640 Nathan Lasnoski It means that every voting experience is a great experience for you on the water and off the water. 0:40:0.890 –> 0:40:15.50 Nathan Lasnoski When I go to put my boat in the water, I want to know that the batteries charged up that I’m going out on a lake that is going to be appropriate for my boat, that the ballast is set properly, that when I left it out there in the yacht harbor that no one got on it cause it was geofence. 0:40:15.60 –> 0:40:23.850 Nathan Lasnoski Like all these things about the actual product itself, if there’s something wrong with it, maybe I need to put some oil in the boat. 0:40:23.860 –> 0:40:25.190 Nathan Lasnoski I need to know what kind of oil they use. 0:40:25.200 –> 0:40:29.230 Nathan Lasnoski I can quickly answer that question, maybe I wanna add on product what kind of product should I use? 0:40:29.300 –> 0:40:31.420 Nathan Lasnoski All this is a digital bonding experience. 0:40:31.910 –> 0:40:38.610 Nathan Lasnoski They even have this relationship with uh, with a like rental bullet kind of organization that they own on. 0:40:38.620 –> 0:40:40.170 Nathan Lasnoski They’re less you like as a boat owner. 0:40:40.180 –> 0:40:44.260 Nathan Lasnoski Start using those those boats and be able to like go out for the day. 0:40:44.490 –> 0:40:47.620 Nathan Lasnoski This idea of being a leader in digital boating, it’s disruptive to the market, right? 0:40:47.630 –> 0:40:56.630 Nathan Lasnoski It’s this idea of I have this experience, this relationship on a continuous basis with the company, not just like I bought this boat, you know, ten years ago. 0:40:56.640 –> 0:40:57.940 Nathan Lasnoski And I never think about it again. 0:40:58.150 –> 0:41:0.530 Nathan Lasnoski I just and then when it breaks, I don’t know how to fix it. 0:41:0.650 –> 0:41:3.740 Nathan Lasnoski This is really their mission is to be the leader in digital boating. 0:41:3.750 –> 0:41:6.500 Nathan Lasnoski So AI is it imperative part of this picture? 0:41:8.310 –> 0:41:17.590 Nathan Lasnoski OK, everything I’ve talked about so far has been like if it’s engaging the physical world, it’s like a digital relationship to the physical world. 0:41:18.700 –> 0:41:20.360 Nathan Lasnoski All these are things you can do now. 0:41:20.430 –> 0:41:26.40 Nathan Lasnoski These aren’t new that these they’re new things, but like they aren’t things that you have to wait till tomorrow to do something. 0:41:26.50 –> 0:41:30.920 Nathan Lasnoski I’m gonna show you now is kind of what’s coming for some of your factory environments. 0:41:31.110 –> 0:41:34.420 Nathan Lasnoski So you already have probably robotics in your factory. 0:41:34.910 –> 0:41:47.460 Nathan Lasnoski Automation has been around for some time, but they do very specific tasks, so if you’re in a production facility for bottles you have about machine that just like it runs those things like nobody’s business, right? 0:41:47.710 –> 0:41:55.0 Nathan Lasnoski If you’re in producing paper it, it’s just like flying through the machines machinery just in general has been around for thousands of years. 0:41:55.310 –> 0:41:59.190 Nathan Lasnoski This is now an example, but of generalizable automation. 0:41:59.390 –> 0:42:15.120 Nathan Lasnoski So this example from Tesla, there’s other vendors doing the same thing that are building generalizable ability for AI agents in physical settings to do tasks that are not highly automated. 0:42:15.130 –> 0:42:19.30 Nathan Lasnoski They’re like something you assign a person to do, like you think about these Amazon warehouses. 0:42:19.380 –> 0:42:21.590 Nathan Lasnoski Why do they need so many people in there? 0:42:21.600 –> 0:42:23.760 Nathan Lasnoski Like it’s not very automated. 0:42:23.770 –> 0:42:30.620 Nathan Lasnoski They got people running around like a chicken with their head cut off in that space, but it’s so generalizable that you can’t naturally take a machine to do it. 0:42:30.950 –> 0:42:37.360 Nathan Lasnoski So they have all these people running around now like they’re painless people pretty well and maybe they will for a long time. 0:42:37.450 –> 0:42:47.980 Nathan Lasnoski But if you get to a point where those generalizable tasks can be done by a robot that can fold laundry, a robot that can put things together, or a robot that can sort blocks all these are on YouTube. 0:42:48.130 –> 0:42:59.160 Nathan Lasnoski These are really interesting scenarios where you start to put together generalizable skill generalizable capabilities physically, and then a set of tasks that you could say all right. 0:42:59.170 –> 0:43:0.680 Nathan Lasnoski Like I want you to fold this laundry all the time. 0:43:0.950 –> 0:43:1.830 Nathan Lasnoski I want you to do this task. 0:43:1.840 –> 0:43:3.740 Nathan Lasnoski I want you to sort things between the boxes. 0:43:3.750 –> 0:43:10.610 Nathan Lasnoski I want you to do this screwing of this thing in like these are all things that sometimes humans do that we now might be seeing in our factories. 0:43:10.780 –> 0:43:12.30 Nathan Lasnoski This is not tomorrow. 0:43:12.40 –> 0:43:21.530 Nathan Lasnoski This is something maybe on the horizon that you might see like three or four years from now, but it’s still something very interesting for you to be thinking about in terms of the job transformations. 0:43:21.920 –> 0:43:28.410 Nathan Lasnoski OK, so all of this makes you think, what do I do from a human first AI strategy? 0:43:28.420 –> 0:43:29.850 Nathan Lasnoski At least I hope you think that way. 0:43:30.110 –> 0:43:36.240 Nathan Lasnoski And for me, that’s the number one thing that organization should be thinking about in the context of how they execute on their mission. 0:43:36.250 –> 0:43:37.880 Nathan Lasnoski I have a mission that exists in the world. 0:43:38.70 –> 0:43:51.900 Nathan Lasnoski I want to bring that mission to as many people as possible, meaning I’m driving my value, but I’m also thinking about how do I think about that value in the context of the people that actually provide it, and how do I help them with their skills? 0:43:52.290 –> 0:43:54.360 Nathan Lasnoski So what we see is that there’s two lanes. 0:43:55.540 –> 0:43:59.490 Nathan Lasnoski The 1st lane is this idea of mission driven, which we talked about. 0:43:59.630 –> 0:44:14.910 Nathan Lasnoski Uh talked about earlier today, so this idea of, like things that truly are tied to the mission of your business in the context of the work that you actually do, customer service building, my thing, inventory, supply chain planning, these are driving what you do mission driven factory. 0:44:14.970 –> 0:44:26.310 Nathan Lasnoski 3 second thing is that commodity lane, everyone in your business, everyone in your business is using AI to be more and that’s why both of these are united by a human first AI strategy. 0:44:26.510 –> 0:44:31.450 Nathan Lasnoski All of this needs to be done in the context of your executive team realizing this is not an IT project. 0:44:31.460 –> 0:44:39.370 Nathan Lasnoski This is a business project or a set of business initiatives that are associated with the entire employee base in your organization. 0:44:40.140 –> 0:44:44.390 Nathan Lasnoski When you think about the roles that exist, this is what’s gonna happen. 0:44:44.480 –> 0:44:49.720 Nathan Lasnoski You’re gonna see this existing state of repetitive work and creative work flip on its head. 0:44:49.790 –> 0:44:53.340 Nathan Lasnoski So most of us, I don’t care what role you have. 0:44:53.670 –> 0:44:58.980 Nathan Lasnoski Most of us come to work every day and we have these things that we can do in our sleep and we just do them. 0:44:58.990 –> 0:45:0.300 Nathan Lasnoski They’re tasks we have to do. 0:45:1.130 –> 0:45:6.200 Nathan Lasnoski For me, it’s like putting a presentation together or creating a statement of work responding to customer. 0:45:6.270 –> 0:45:10.140 Nathan Lasnoski I just know how to do these things and they’re like tasks that I take care. 0:45:10.190 –> 0:45:13.80 Nathan Lasnoski Maybe someone processes, AR automation all day long. 0:45:13.90 –> 0:45:22.340 Nathan Lasnoski Or they input customer, they deliver customer products, they or they put things together or they do quotes like all these are repetitive work that you do. 0:45:22.350 –> 0:45:23.850 Nathan Lasnoski And this is like sliver, what you do? 0:45:23.860 –> 0:45:35.290 Nathan Lasnoski That’s just creative work, where you truly have to like turn on the best version of yourself and ideate and create and a spate a way that you could kind of identify that is like, what does a person do when there’s nothing on their calendar? 0:45:35.970 –> 0:45:40.460 Nathan Lasnoski Like we always meetings, right when you have a free black, how do you use that time? 0:45:40.730 –> 0:45:42.500 Nathan Lasnoski That’s this right? 0:45:42.510 –> 0:45:44.920 Nathan Lasnoski The creative work that’s that free block. 0:45:45.190 –> 0:45:51.250 Nathan Lasnoski How many of us are prepared to use that time effectively? 0:45:51.460 –> 0:45:58.550 Nathan Lasnoski How many of us are prepared for that free block to be a bigger chunk of their time and to use that to drive more outcomes? 0:45:58.560 –> 0:46:2.410 Nathan Lasnoski Working on the business, not just in the business, that’s the transition. 0:46:2.420 –> 0:46:3.630 Nathan Lasnoski That’s scary for people. 0:46:3.920 –> 0:46:13.40 Nathan Lasnoski We had a group at our office last night was an executive AI session on Manufacturing lot of manufacturers and the person that was speaking, he said. 0:46:13.970 –> 0:46:24.260 Nathan Lasnoski He has to be super proactive about people realizing that there’s a journey for each of them, and it’s very personal to them because they do this task. 0:46:24.430 –> 0:46:25.540 Nathan Lasnoski It’s connected to them. 0:46:25.550 –> 0:46:31.760 Nathan Lasnoski It’s connected to who they are and when they see that task be removed, they’re like, what does that mean for me? 0:46:31.770 –> 0:46:33.840 Nathan Lasnoski I thought about that application. 0:46:33.850 –> 0:46:36.100 Nathan Lasnoski I brought that application in the into our business. 0:46:36.450 –> 0:46:39.500 Nathan Lasnoski I do that job for the last 20 years. 0:46:39.510 –> 0:46:41.980 Nathan Lasnoski Like what does it mean for me, the winners? 0:46:41.990 –> 0:46:47.10 Nathan Lasnoski The people that are going to be engage, you’re gonna think this is opportunity for me to create more. 0:46:47.360 –> 0:46:50.470 Nathan Lasnoski But some people are gonna be like this is the Maslow’s hierarchy of needs. 0:46:50.480 –> 0:46:51.750 Nathan Lasnoski You’re taking something away? 0:46:51.820 –> 0:46:58.810 Nathan Lasnoski I’m not sure what to do with that problem, so this transition is important and it’s not just an IT transition, it’s a transition for every person. 0:46:59.280 –> 0:47:0.430 Nathan Lasnoski So how do you get started? 0:47:0.440 –> 0:47:1.650 Nathan Lasnoski What do we go from here? 0:47:1.990 –> 0:47:3.610 Nathan Lasnoski Do I learn data science curriculum? 0:47:3.620 –> 0:47:4.990 Nathan Lasnoski Does everybody Learn AI? 0:47:5.360 –> 0:47:6.350 Nathan Lasnoski The answer to that is no. 0:47:6.680 –> 0:47:17.30 Nathan Lasnoski Most of this, like this data science itself, will impact less than .1% of your staff, if even that much data science is not something that everyone’s gonna be doing. 0:47:17.240 –> 0:47:21.150 Nathan Lasnoski Data science is about, ah, certain set of people. 0:47:21.160 –> 0:47:23.630 Nathan Lasnoski AI upskilling is not about data science. 0:47:23.720 –> 0:47:27.810 Nathan Lasnoski AI upskilling is about everyone using an AI assistant. 0:47:27.900 –> 0:47:35.750 Nathan Lasnoski Being a practitioner of assets and AI space, not just a builder, adoption of skills, not roles. 0:47:35.760 –> 0:47:45.750 Nathan Lasnoski So taking people that are in certain functions, adopting skills to those functions, not building a new role, that they’re then doing, it’s not limited to technical skills. 0:47:45.820 –> 0:47:54.50 Nathan Lasnoski So this idea of, like colleges are building these like AI certificates and so on like it’s it’s I mean well good idea. 0:47:54.60 –> 0:47:54.550 Nathan Lasnoski OK. 0:47:54.620 –> 0:47:59.610 Nathan Lasnoski But like, that’s not what like the 95% of our humanity is gonna need. 0:47:59.720 –> 0:48:9.390 Nathan Lasnoski There’s this whole chunk of people that need to be more unlock creative skills and other skills they haven’t enabled in a while, which is this growth mindset kind of space. 0:48:9.820 –> 0:48:17.810 Nathan Lasnoski So when you look at those roles in the context of the overall stack, you have data engineers that prepare the data for the environment. 0:48:18.280 –> 0:48:22.550 Nathan Lasnoski You have data scientists that build the trusted AI models. 0:48:22.900 –> 0:48:30.910 Nathan Lasnoski You have AI engineers that use those trusted models, a foundational models to put the like pieces together to be able to accomplish the goals. 0:48:31.100 –> 0:48:38.70 Nathan Lasnoski There might be a little bit like here which is like people that can put together Lego blocks but not build Lego blocks. 0:48:38.300 –> 0:48:43.930 Nathan Lasnoski So there’s kind of a diversity of what AI engineers might contain, but then there’s truly the AI practitioner. 0:48:44.40 –> 0:48:48.990 Nathan Lasnoski And they create value in their everyday work by using AI their expert in the business. 0:48:49.300 –> 0:48:56.690 Nathan Lasnoski They’re not necessarily expert at building AI systems, so AI practitioners is where most of you are going to find your career. 0:48:56.700 –> 0:49:2.290 Nathan Lasnoski Continuing to move into where many of you will then see like these other tasks to be able to prepare for those environments. 0:49:3.290 –> 0:49:9.740 Nathan Lasnoski So all of that kind of fits into a picture and somehow we’ve gone through this in 50 minutes. 0:49:9.750 –> 0:49:12.310 Nathan Lasnoski So I’m going to have some time for questions, which is fantastic. 0:49:12.380 –> 0:49:15.570 Nathan Lasnoski So before you go, I want you to see these things. 0:49:16.0 –> 0:49:18.510 Nathan Lasnoski These are what we would suggest for next steps. 0:49:18.740 –> 0:49:23.230 Nathan Lasnoski So you already have in your chat and Lync. 0:49:23.280 –> 0:49:24.530 Nathan Lasnoski I think you have it in your chat. 0:49:24.600 –> 0:49:25.590 Nathan Lasnoski You should have a Lync. 0:49:25.20 –> 0:49:26.80 Amy Cousland I’m about to do it now. 0:49:26.90 –> 0:49:27.50 Amy Cousland I’ll set it out in a second. 0:49:26.120 –> 0:49:27.190 Nathan Lasnoski You’re about to do it. 0:49:27.200 –> 0:49:27.620 Nathan Lasnoski Cool. 0:49:27.630 –> 0:49:28.150 Nathan Lasnoski Thank you. 0:49:28.600 –> 0:49:35.170 Nathan Lasnoski OK, so you should have in your chat a link to fill out a survey, and the goal of that is two things. 0:49:35.180 –> 0:49:39.690 Nathan Lasnoski First, I want you to know if you like this and found an interesting like I always looking for feedback. 0:49:39.760 –> 0:49:43.280 Nathan Lasnoski This is something I really enjoy doing, but I wanna do it the best possible. 0:49:43.290 –> 0:49:48.210 Nathan Lasnoski I can’t, and I’m always looking to be able to improve it and provide more value for these kinds of conversations. 0:49:48.230 –> 0:49:56.340 Nathan Lasnoski We’re gonna do a manufacturing version two and version three later in the year, so I want to make sure that if you have feedback, things you love, things you didn’t love, you put it in there. 0:49:56.390 –> 0:50:4.600 Nathan Lasnoski 2nd, we want you to be able to make this real in your business, so AI and Copal and visioning workshops is something we’re doing literally all the time. 0:50:4.610 –> 0:50:7.720 Nathan Lasnoski Like I think I had like 5 this week with executive teams. 0:50:7.730 –> 0:50:16.490 Nathan Lasnoski So like CEO, CFO, COO, kind of conversations on so envisioning workshops, we can do Microsoft funded use case validation. 0:50:16.500 –> 0:50:19.600 Nathan Lasnoski So if you already have something you’re looking at, you’re like I already know what I want to do. 0:50:20.610 –> 0:50:22.260 Nathan Lasnoski Here’s the thing, I’ve already got approved. 0:50:22.270 –> 0:50:25.770 Nathan Lasnoski The executive team wants to do it like we’re we’re often running, but I need someone to help me. 0:50:26.30 –> 0:50:29.890 Nathan Lasnoski Like, that’s something that we can do to help evaluate that use case for free. 0:50:30.930 –> 0:50:40.340 Nathan Lasnoski And then if you’re going down the copilot, copilot, space or the AI space, and you’re at that point where, like you already know, as an executive team, that is important and they’re aligned. 0:50:40.430 –> 0:50:41.920 Nathan Lasnoski And they have the art of the possible. 0:50:42.30 –> 0:50:45.340 Nathan Lasnoski But you want to start figuring out the list of ideas we’ve done sessions. 0:50:45.350 –> 0:50:56.980 Nathan Lasnoski Up to 100 people to bring, like, broaden the tent in the organization, across sales and marketing and all these different groups bring them together to be able to have a conversation, to be able to figure out what the right ideas are. 0:50:57.170 –> 0:50:58.860 Nathan Lasnoski And that’s been super powerful. 0:50:59.30 –> 0:51:1.740 Nathan Lasnoski So any of these are things we love to do with you. 0:51:1.750 –> 0:51:5.930 Nathan Lasnoski And please check one and just love to have the next step conversation with anyone of you. 0:51:6.260 –> 0:51:10.670 Nathan Lasnoski OK, so let’s go into some questions. 0:51:11.180 –> 0:51:14.860 Nathan Lasnoski So, or there’s a there’s a very long. 0:51:14.20 –> 0:51:25.230 Amy Cousland And you, so you got a question from Paul here about why do people assume the human form factor is the most efficient form for completing tasks that we had somebody that uses AI to answer that for him? 0:51:25.240 –> 0:51:26.610 Amy Cousland So if you have, if you can weigh in on that. 0:51:27.780 –> 0:51:29.70 Nathan Lasnoski That’s a good question. 0:51:29.80 –> 0:51:31.710 Nathan Lasnoski Why is it the most effective human uh form factor? 0:51:31.720 –> 0:51:32.650 Nathan Lasnoski It might not be. 0:51:32.710 –> 0:51:33.750 Nathan Lasnoski That’s a great question. 0:51:34.620 –> 0:51:41.870 Nathan Lasnoski Maybe it’s because we have a generalizable ability that machines don’t, and we’re sort of copying the human form factor in that context. 0:51:42.310 –> 0:51:47.910 Nathan Lasnoski That’s my guess, that generalizability of what we have the ability to do is like a step. 0:51:48.600 –> 0:51:54.850 Nathan Lasnoski Maybe they’ll start replacing those like finger arms with something that’s more effective tool for some of the generalizable tasks they’re doing. 0:51:55.300 –> 0:51:57.190 Nathan Lasnoski Umm, that’s a little scary. 0:51:57.200 –> 0:52:2.60 Nathan Lasnoski Like kinda reminds me of the team one T800 or something but umm I don’t know. 0:52:2.70 –> 0:52:3.250 Nathan Lasnoski I don’t better answer to that. 0:52:3.350 –> 0:52:4.200 Nathan Lasnoski It’s a good question. 0:52:4.590 –> 0:52:7.470 Nathan Lasnoski Nice use of chat, GPT, OK. 0:52:9.330 –> 0:52:18.80 Nathan Lasnoski Thank you and answered generated from data accumulated from humans, I wonder if we can ever imagine outside that box stated cumulative be quite fun from humans. 0:52:18.960 –> 0:52:21.670 Nathan Lasnoski Answer generating from data accumulated by humans. 0:52:21.680 –> 0:52:24.220 Nathan Lasnoski I wonder if we can ever imagine outside of that box. 0:52:24.290 –> 0:52:26.400 Nathan Lasnoski This I think we can. 0:52:26.570 –> 0:52:31.90 Nathan Lasnoski I mean, think about the, I mean maybe not the. 0:52:31.550 –> 0:52:39.50 Nathan Lasnoski Uh, well, this step back for a second like Chad GPT type answers. 0:52:39.60 –> 0:52:45.470 Nathan Lasnoski Yes, very much grounded in the context of what it’s learned before, and there’s not a lot of relearning happening within that space. 0:52:45.480 –> 0:52:49.250 Nathan Lasnoski So the idea that like it’s learning new things, not so right. 0:52:49.260 –> 0:52:51.470 Nathan Lasnoski It’s like kind of boxed in in that space. 0:52:51.580 –> 0:52:55.170 Nathan Lasnoski Humans like you could say, is the human context boxed in? 0:52:55.180 –> 0:52:56.800 Nathan Lasnoski Sure, it’s boxed in by our senses. 0:52:56.810 –> 0:53:1.670 Nathan Lasnoski Is boxed in by our what’s written on our hearts so that it’s boxed in by. 0:53:1.740 –> 0:53:4.250 Nathan Lasnoski What’s the world around us and the interactions that we have? 0:53:4.560 –> 0:53:8.550 Nathan Lasnoski That’s a pretty broad, automatable summative space, right? 0:53:8.670 –> 0:53:11.300 Nathan Lasnoski So I don’t know if we’re necessarily boxed in. 0:53:11.310 –> 0:53:15.220 Nathan Lasnoski Well, AI agents have the ability to do that same thing once they can learn continuously. 0:53:15.670 –> 0:53:23.450 Nathan Lasnoski We definitely seen that AI agents when put in spots like you put it to AI agents in a game, you give them the rules. 0:53:23.460 –> 0:53:24.290 Nathan Lasnoski You tell them to learn it. 0:53:25.270 –> 0:53:26.280 Nathan Lasnoski They’ll run a million times. 0:53:26.290 –> 0:53:34.640 Nathan Lasnoski It gets better as it goes, so the ability to like find things from the where to surround you is is something that I agents may possess the ability to do as well. 0:53:34.810 –> 0:53:38.580 Nathan Lasnoski So I think we’ll get to an interesting spot where AI agents can do some of those similar things. 0:53:38.590 –> 0:53:41.760 Nathan Lasnoski And that’s, I think, the fascinating future, but not something we have right now. 0:53:42.420 –> 0:53:43.180 Nathan Lasnoski Easy question. 0:53:43.190 –> 0:53:44.80 Nathan Lasnoski Do you need Windows 11? 0:53:44.90 –> 0:53:45.370 Nathan Lasnoski Use copilot within the OS. 0:53:46.570 –> 0:53:48.40 Nathan Lasnoski I don’t know if that’s an easy question. 0:53:48.50 –> 0:53:50.60 Nathan Lasnoski Actually, I think the answer to that is yes. 0:53:50.150 –> 0:53:51.10 Nathan Lasnoski Thanks Chris Blackburn. 0:53:53.950 –> 0:53:54.870 Nathan Lasnoski I actually prefer. 0:53:54.880 –> 0:53:58.150 Nathan Lasnoski I’ll just tell you I prefer M365 copilot. 0:53:58.160 –> 0:54:16.0 Nathan Lasnoski It’s much more it’s cause it’s grounded in the capabilities of the old 365 tenant and I also prefer the copilot on the ground, the ground and copilot in the web over the Windows 11 copilot at the moment, just as a data point when dealing with security. 0:54:16.10 –> 0:54:22.120 Nathan Lasnoski How do you weigh using AI with our own data versus pulling in data from outside sources for solutions? 0:54:22.170 –> 0:54:23.240 Nathan Lasnoski Great question. 0:54:23.890 –> 0:54:28.160 Nathan Lasnoski The challenge with the outside data is someone could change the source of they had outside data. 0:54:28.170 –> 0:54:33.550 Nathan Lasnoski So you actually something we’ve been seeing companies try to hack AI solutions by changing the source data. 0:54:33.560 –> 0:54:40.160 Nathan Lasnoski You probably saw some examples of like Amazon where like you can see that it’s messed that up and other use cases. 0:54:40.390 –> 0:54:46.290 Nathan Lasnoski So the number one thing about building AI solution is you have to know how you’re using the data and where it came from. 0:54:46.440 –> 0:54:50.510 Nathan Lasnoski So most AI solutions were using our grounded only in internal data. 0:54:51.400 –> 0:55:5.510 Nathan Lasnoski Unless you’re talking about like a copilot where like, you’re just sort of like 365 copilot, where you maybe combining it with things you can get from the web and it’s like more of like, I’m giving information back, it’s better than Google search, but it’s not like a definitive answer. 0:55:5.600 –> 0:55:11.830 Nathan Lasnoski That’s probably OK to go for the web, but most of the solutions we’re building are grounded in data that’s owned by the company. 0:55:12.150 –> 0:55:15.670 Nathan Lasnoski Customer service chat bots for walking through a process. 0:55:16.280 –> 0:55:19.190 Nathan Lasnoski Internal chat bots for specific HR policies. 0:55:19.200 –> 0:55:21.10 Nathan Lasnoski It’s all grounded in private data. 0:55:21.160 –> 0:55:27.250 Nathan Lasnoski I generally wouldn’t like point that at the Internet and say like answer this question for me, if it’s tied to something that’s critical. 0:55:27.310 –> 0:55:29.230 Nathan Lasnoski Where I have to have precision and accuracy. 0:55:29.330 –> 0:55:32.690 Nathan Lasnoski You want that to be grounded in private data, so that’s a good question. 0:55:33.120 –> 0:55:35.280 Nathan Lasnoski Definitely a focus on using private data. 0:55:38.690 –> 0:55:39.80 Nathan Lasnoski OK. 0:55:39.90 –> 0:55:39.730 Nathan Lasnoski Any other questions? 0:55:46.0 –> 0:55:46.550 Nathan Lasnoski Thanks Chris. 0:55:49.720 –> 0:55:54.800 Nathan Lasnoski When that limit you to being stuck in the, that’s the way we’ve always done it mentality. 0:55:56.660 –> 0:55:58.320 Nathan Lasnoski This was it depends upon the use case. 0:55:58.790 –> 0:56:8.40 Nathan Lasnoski UM, if you are doing customer service and you wanna customer service bot to answer the question specifically for those reps or for your end customers. 0:56:8.50 –> 0:56:10.750 Nathan Lasnoski So that replaces the right part or puts the right oil and engine. 0:56:11.840 –> 0:56:15.290 Nathan Lasnoski No, I actually don’t want it to go outside the box like I want it to you. 0:56:15.300 –> 0:56:22.50 Nathan Lasnoski Exactly what I tell it to do, and honestly, that’s probably the majority of scenarios you’ve run into in the creative space. 0:56:22.60 –> 0:56:22.890 Nathan Lasnoski Totally not right. 0:56:22.900 –> 0:56:33.370 Nathan Lasnoski You want it to expand to a variety of things that are outside the box, and in that case you probably would give it external sources, but that’s not one where you’re going toward accuracy and precision. 0:56:33.380 –> 0:56:38.800 Nathan Lasnoski That’s one you’re going toward creativity, which is totally different than the other use case. 0:56:39.800 –> 0:56:47.810 Nathan Lasnoski So good question though I think that’s something worth like, I think that’s like a purpose driven conversation over beers conversation we should have. 0:56:53.130 –> 0:56:54.60 Nathan Lasnoski Alright everybody. 0:56:54.150 –> 0:56:54.740 Nathan Lasnoski Thank you. 0:56:54.810 –> 0:56:55.580 Nathan Lasnoski This was fun. 0:56:55.860 –> 0:56:57.820 Nathan Lasnoski I I hope this was interesting to you. 0:56:57.930 –> 0:57:0.620 Nathan Lasnoski I really enjoyed presenting it every time we do this. 0:57:0.630 –> 0:57:3.0 Nathan Lasnoski It’s more cool because there’s more more use cases. 0:57:3.10 –> 0:57:5.570 Nathan Lasnoski I hope you have a great day and thank you for joining us today.